μPyCSP - Two approaches to implementing CSP-like concurrency modeling in microcontrollers.

Abstract

In this thesis, we have explored the possibility of bringing CSP-like concurrency modeling to an embedded environment. With the growth of IoT and embedded software, many of the hardships that come with concurrent coding in an inherently event-driven environment become more apparent. Convoluted and complex code is often used to handle the many fallacies of concurrency. We are solving this problem by abstracting away the concurrency and providing an easy to use interface, in a growing language; Micropython. Through two separate library implementations using different underlying asynchronous architecture, harvesting the potentials and of multi-threading and coroutine based CSP-processes.

We have shown through measurements and example implementations that both implementations provide viable performance, and uncovered the different advantages and disadvantages to each approach.